For a library, GDB does only see source in header files, whereas source in .cpp files is not available

Question

I am using gdb to debug a program compiled against a static library. Both the library and the program were compiled with debug symbols. For this library, gdb does only see the source code implemented in .h files and not the source code implemented in .cpp files. I wonder why?

• For example

  (gdb) info sources
  

  only lists header files from that library and no .cpp files.

• I am able to set breakpoint to some function that was implemented in a .cpp file. For example gdb understands this command:

  (gdb) break orgQhull::Qhull::runQhull,
  

  it is, however, not possible to debug the code inside the body of that function. (It is only possible to step through individual instructions on the assembler level using the stepi command.) On the other hand, the debuger is able to see the code of functions implemented in header files.

• The backtrace command reports that there is "No symbol table available" for the runQhull function mentioned earlier:

  (gdb) backtrace full
  
  #0  0x00000000004d361a in orgQhull::Qhull::runQhull(orgQhull::RboxPoints const&, char const*) ()
  No symbol table info available.
  #1  0x000000000049573a in ConvexHull::ConvexHull (this=0x7fffffffcb70, tree=..., begin=..., end=...)
      at /home/filip/fel/dp/sandbox/src/populating_octomap/include/ConvexHull.h:40
  

What could be the reason for this? Any suggestions are welcome!

Some details:

• The library in question is Qhull's libqhullcpp.a with sources at git://gitorious.org/qhull/qhull.git
• System: Ubuntu 12.04.2 LTS with GNU/Linux 3.2.0-40-generic #64-Ubuntu SMP x86_64
• gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3
• GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2.1) 7.4-2012.04

Answer 1

What could be the reason for this?

Either

• you've built the library without -g flag, or
• you've stripped the library after it was built (this is very unlikely).

Check your library build log. You believe you built it with -g, but you are very likely mistaken.